© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 92535
    Prev
    Next
    INDEX
    Explanations

    Your

    np_max-act · gemini-2.0-flash

    The neuron identifies polite email‐reply phrases—especially tokens around thanking someone for their email (e.g. “thank,” “for,” “your,” “email,” etc.).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.11.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    -tag
    -0.06
    主任
    -0.06
    _tok
    -0.06
    ोल
    -0.06
    	constexpr
    -0.06
    сыл
    -0.06
    บาง
    -0.06
    เร
    -0.06
    nesc
    -0.06
    过程
    -0.06
    POSITIVE LOGITS
     chambre
    0.07
    \v
    0.07
    Ậ
    0.07
    ikel
    0.06
    Vous
    0.06
     legit
    0.06
     sü
    0.06
    Ö
    0.06
     forget
    0.06
    ��
    0.06
    Activations Density 0.021%

    No Known Activations